28o SILVER ALLOYS
two tests should have shining, hemispherical upper surfaces and opaque,
white lower ones, and should differ in weight by a few milligrams at most.
The method of cupellation, largely used for the analysis of argentiferous
minerals and for the control of intermediate products in the extraction of silver,
as well as for the by-products and for low-grade alloys, is not advisable for
analysing ordinary jewellery, coinage, etc., since, however carefully it is carried
out, the results obtained are not always concordant and never very exact, being
mostly somewhat low. The errors are due principally to volatilisation of the
silver and absorption by the cupel. Tables showing the corrections to be applied
have been prepared, but the best method of determining such corrections is
to carry out a check determination, at the same time and under the same con-
ditions, with pure silver and pure copper in approximately the same proportions
as in the sample.
As regards the influence of extraneous metals on the results, it should be
borne in mind that gold and platinum remain with the silver and increase its weight.
They may be detected by treating the button with nitric acid and examining the
black powder remaining undissolved. In low proportions, arsenic, tin, antimony,
bismuth, iron, nickel and cobalt do not interfere appreciably with the operation.
(6) VOLHARD'S METHOD.1 This method consists in precipitating the
silver in nitric acid solution with standard ammonium thiocyanate solution
in presence of ferric sulphate as indicator. As soon as the precipitation of
the silver is complete, the thiocyanate reacts with the ferric salt and gives
a persistent red coloration, which marks the end of 4he reaction.
Reagents, (i) A solution containing 3-1-3-2 grams of ammonium
thiocyanate, free from chlorides, per litre.
\ (2) Cold saturated ferric ammonium alum solution free from chlorides,
and treated with a little nitric acid to destroy the brown colour ; the same
amount (2-3 c.c.) is used in each titration.
(3) Pure silver (fine silver}. This may be obtained in foil from reputable
Where many tests are made the fine silver is prepared in the laboratory
from the silver chloride residues from Gay-Lussac's method or from silver
chloride precipitated from the silver nitrate solutions obtained in the quctrta-
tion of gold (see later). Failing these, commercial silver of 999 fineness is
dissolved in nitric acid, allowed to stand for some days in the dark, filtered
to remove traces of undissolved gold and the silver precipitated with a
slight excess of dilute hydrochloric acid.
In whatever way obtained, the silver chloride is washed free from acid,
dissolved in ammonia, left for some days and then filtered, the clear solution
being made acid with dilute hydrochloric acid to precipitate the silver
chloride, which is again washed free from acidity and redissolved in ammonia.
After standing for some time, the solution is filtered, treated with sodium
hydroxide in the proportion of 750 grams per 1000 grams of the chloride
and heated to boiling, 150-200 grams of pure, powdered glucose being added
in small amounts and the liquid kept boiling for about 30 minutes. The
spongy silver thus obtained is pumped off, thoroughly washed, dried and
fused in a refractory crucible with a little nitre and borax.
When cold, the crucible is broken and the metallic button washed with
1 Also known as the Charpentier-Volhard method.icks round the edges of the cupel, whilst, if too high,